Minutes of the Physics Commissioning Meeting June 30, 1999
Michael Schmelling
==========================================================
The focus of the meeting was the on-line data quality monitoring
for the coming running period. After a presentation by Maarten
Bruinsma about the status of the monitoring and data-quality
framework, the various subsystems gave a brief review of their
monitoring status.
Monitoring framework: Maarten Bruinsma/Dorothea Samtleben
=====================
A first version of the Data-Quality Monitor (DQM) by Maarten is
ready, as is the L4-monitoring framework provided by Dorothea.
The monitor is a simple RHP-client which in the same way as it
subscribes to the L4 monitoring can also subscribe to other RHP
servers. Monitoring performed on DMON-level thus can be integrated
in a transparent way.
Subsystem reports:
==================
Target: Youri Vassiliev
-------
Three of the four monitoring quantities dicussed on June 16
are available in the target monitoring. Integration into RHP
is still pending.
VDS: Martin Braeuer
----
Using the framework by Dorothea, a first set of L4-monitoring
histograms has been implemented for the monitoring of
ocupancies, cluster charges, signal/noise, tracking residuals.
DMON monitoring exists for the VDS, but still has to be
integrated into the DQM.
Inner Tracker: Wolfgang Gradl
--------------
Inner tracker monitoring, partially based on VDS software, exists
on DMON level for basic quantities like BX-contributions,
signal/noise and occupancies.
Outer Tracker: Uli Uwer
--------------
Low-Level checks that are needed for chamber commissioning exist
at DMON level. A reduced set of basic histograms which allow per
superlayer checks of data integrity and quality are currently
being setup. High level monitoring on track level, which requires
L4 still requires a substantial amount of work.
RICH: Denis Dujmic
-----
The RICH monitoring, using DMON and on L4, is organized hierarchically
with basic quantities being output roughly every 1000 events, more
sophisticated ones written every 5000 events and high level
characteristics of the detector peformance stored on per run basis.
Most of the coding has been done.
HIPT: Vladimir Popov
-----
Occupancies are monitored with DMON, correlations with other
detectors are looked at at L4. The software exists.
TRD: -- not represented --
----
ECAL: Sergey Shuvalov
-----
ECAL monitoring is integrated into the general ECAL slow control
using different data paths and some specialized software to handle
large histogram objects. The code is almost ready. Integration into
the DQM would require putting summary information into RHP-objects.
MUON: Marc Buchler (presented by Boris Fomynikh)
-----
Monitoring code for L4 to a large extent has been written, looking
at target spots, efficiencies or track residual. On DMON level
synchronization information (occupancy vs BX) exist, additional
quantities can be added if needed. Figure-of-merits are not yet
defined.
DAQ: Federico Sanchez
----
Since DAQ is a subsystem of a different quality as the individual
subdetectors, monitoring also has a different aspects. Information
that could go to the DQM is related to the FCS (fast control system)
like for example trigger rejects. Monitoring exists, but would
have to be integrated into the DQM.
SLT: Mike Medinnis
----
SLT like DAQ is not a data source and thus also provides monitoring
of a different quality. As it sensitively depends on the other
components working, monitoring to some extent is redundant. Still,
many quantities merit being looked at and also transferred to the DQM.
General discussion
==================
1) DMON
-------
Most groups already have both low-level and high-level monitoring,
the former running at DMON-level or a dedicated data channel, the
latter one being installed at L4. Integration of both into the DQM
should be straightforward.
The DMON Low-level monitoring has advantage that it needs much less
infrastructure than L4-monitoring and in addition allows to implement
additional subdetector specific diagnostics without using FARM
resources. Originally only envisaged as a temporary monitoring option,
it should be decided soon whether to keep it also in the future.
2) Duplicating information
--------------------------
Some quantities can be monitored both at DMON level and at L4. The
groups are asked to share the monitoring such that resources are
used as effiently as possible.
3) Bandwidth issues
-------------------
The L4-monitoring consumes memory, CPU-power and using RHP also
network bandwidth. The actual requirements by the individual groups
are not yet clear, but could be a potential bottleneck. Estimates
about the necessary resources are needed as soon possible.
4) Software project DQP
-----------------------
As the data quality monitoring becomes a central item for the
Hera-B commissioning and running it also should become a distinct
software project. One should also discuss integrating into it
the hbdisplay-related monitoring activities. (Note that the
DQM-software has been built using of hbdisplay software.)
-------------------------------------------------------------------------
Michael Schmelling
Max-Planck-Institut fuer Kernphysik
Mailing Address: e-mail: Michael.S...@mpi-hd.mpg.de
Postfach 103980 Phone : [49](6221)516-511
D-69029 Heidelberg Fax : [49](6221)516-603
-------------------------------------------------------------------------
Maarten Bruinsma
Dear Collegues,
I am happy to announce the first beta (gamma?) release of the
monitor client program to be used for Data Quality Monitoring.
Unfortunately I have not been able to debug it as much as I would have liked
and therefore bugs can be expected. However, in its present status
it is usable. Furthermore there exists no documentation yet.
You can invoke the client program as follows:
1) Log on to one of the hb-sun's as hbshift
2) >cd DQM
3) >DQMon
4) Resize the control panel ('feature')
During startup all existing rhp servers are scanned for
and the histograms they serve are identified. One can
subscribe to a histogram by double-clicking the icon
in the list-tree below. An "S" will appear in the
icon indicating a succesful subscription. From then
on the histogram gets automatically updated via sockets.
One can unsubscribe by pressing the right mouse button once
on the icon.
Macros that give summarized and layouted information can
in the future be invoked by pressing the appropriate button. This
is in a development phase and needs input from the detector groups.
On the Software-project DQP:
1) The farm server software is maintained by Dorothea. If you want
higher level histograms served to and checked by the shiftcrew
you should talk with her about the integration into the existing
programs.
2) I am very happy to annouce that I have found Maaijke. willing to be
responsible for the client software. She has been introduced to
the code. Questions should therefore be addressed to her.
One of the most important features still missing are the
macros to be invoked by pressing buttons. The detector groups
should think and decide what they want the shiftcrew
to monitor. This should then be communicated to Maaijke.
As soon as I get back in three weeks I will assist her in finalizing
the monitor program. I hope in the meantime she can count on
your help.
Maarten and Maaijke.
> Monitoring framework: Maarten Bruinsma/Dorothea Samtleben
> =====================
> A first version of the Data-Quality Monitor (DQM) by Maarten is
> ready, as is the L4-monitoring framework provided by Dorothea.
> The monitor is a simple RHP-client which in the same way as it
> subscribes to the L4 monitoring can also subscribe to other RHP
> servers. Monitoring performed on DMON-level thus can be integrated
> in a transparent way.
>
> 4) Software project DQP